Re: [PATCH]generic usb serial driver
Greg KH <[email protected]>
| Newsgroups | gmane.linux.usb.devel |
|---|---|
| Message-ID | <[email protected]> |
On Fri, Mar 30, 2007 at 10:38:01AM -0700, Pete Zaitcev wrote: > On Fri, 30 Mar 2007 11:17:28 +0200, Oliver Neukum <[email protected]> wrote: > > > This can'be solved with a simple spinlock, as wait_event() can sleep. > > That's because wait_event() and its ilk make a bad API. I mentioned it > many times before, and the answer by its pushers was "make the condition > to be a function call and take locks inside it". That's just dumb, > because it works around the silly design. The correct answer is not > to use wait_event ever. Agreed. > > Really somebody went to some effort to find a lockless solution and > > did well, he just didn't figure out of order execution units. You'd destroy > > that. > > That somebody obviously needs a girlfriend. Lockless execution in a > printer driver, that's awesome. > > The bottom line is, there's no excuse for using any kind of smp_mb() > in USB drivers. None at all. When we have 100Gbit/s USB, then perhaps > we can talk about it. Perhaps then, but even then, I doubt it... I really don't want the "normal" driver developer to have to worry about things like smp_mb() in their code. We have a hard enough time dealing with other, more basic race condition issues :) thanks, greg k-h ------------------------------------------------------------------------- Take Surveys. Earn Cash. Influence the Future of IT Join SourceForge.net's Techsay panel and you'll get the chance to share your opinions on IT & business topics through brief surveys-and earn cash http://www.techsay.com/default.php?page=join.php&p=sourceforge&CID=DEVDEV _______________________________________________ [email protected] To unsubscribe, use the last form field at: https://lists.sourceforge.net/lists/listinfo/linux-usb-devel